【摘要】 针对工业过程和实际控制对象的慢时变非线性的特点 ,设计了一种预测模型的单神经元 PI控制器。采用单神经元 PI控制算法与神经网络预测模型相结合的控制策略 ,用 PI控制规律来确定控制器的输出。用一个自适应神经元网络作为非线性系统的预测模型 ,估计下一步的输出值 ;用一个单神经元实现 PI控制来优化下一步的控制。利用 Matlab/Sim ulink工具对 PI控制器和预测模型的单神经元 PI控制器进行比较仿真实验 ,其控制对象为典型的非线性系统。仿真实验表明 :预测模型的单神经元 PI控制器具有结构简单 ,计算速度快 ,鲁棒性好等特点更多还原

【Abstract】 A single neuron PI controller with predictive model is designed according to nonlinear systems of many industry processes and practical plants. The control strategy unites single neuron PI control algorithm and the predictive model based on neural network. The output of the controller is determined by PI algorithm. An adaptive neural network is used to build the predictive model of the nonlinear system which is used to predict the next output of the system. The PI controller that is implemented using a single neuron optimizes next control value. The comparison simulations have been implemented for PI controller and single neuron PI controller with predictive model by Matlab/Simulink tool in the typical nonlinear system. The results of simulation show that the single neuron PI controller with predictive model has simple structure, fast calculation and good robust.更多还原